Indent repairs, Dutchman repairs, and dowelling: how stone actually gets fixed

Repair is a different discipline from cleaning

Everything else in The Grain is, in one way or another, about removing something — soot with DOFF, sulphate crusts with TORC, rust and salts with a poultice. Repair is the opposite problem. Something is missing or broken, and the decision is what, if anything, goes back. The governing principle is the same one that governs every method choice on a Trebor project — minimum intervention — but here it cuts the other way. The question is not how little can be removed. It is how little needs to be added.

That distinction matters because the instinct of an unspecialist contractor, faced with a damaged stone, is usually to make it look new. That is precisely wrong. A repair is successful when it stabilises the stone, reads correctly at a normal viewing distance, and does not introduce a fixing or a material that fails faster than the stone around it. Making a repair invisible at six inches, in raking light, is not the goal. Making it last fifty years without cracking the original stone is.

Reading the damage before touching the stone

The first question on any repair survey is whether the stone is sound behind the visible damage, or whether the loss is a symptom of something still active underneath it. Spalling caused by a one-off impact is a different problem to spalling caused by a corroding fixing that is still corroding. Filling the visible defect while leaving an active cause in place is a repair that is scheduled to fail before it has even cured — which is why a survey on an older building spends as much time on what caused the damage as on the damage itself.

This is also where a repair survey overlaps with what we look for during a facade clean. Rust bleed at a cramp point, discussed in our article on poultice cleaning, is frequently the first visible sign of a fixing that is not just staining the stone but is actively splitting it from within. Cleaning the stain and repairing the stone are two different jobs, but on a building with Victorian ironwork they are very often the same conversation.

Indent repair: cutting out the bad, piecing in the good

An indent repair — new stone let into a socket cut back to sound material — is the method used where the loss is too large or too exposed for a mortar fill to be appropriate: a missing section of moulding, a spalled corner on a quoin, a cracked and displaced piece of a cill or a step. The damaged material is cut back with square, slightly undercut margins, well past the visible extent of the decay and into stone that is genuinely sound, not just stone that looks sound from the front. A repair set into weak or friable material behind the visible face fails at the same rate as the defect it replaced.

The replacement piece is sourced to match the original in colour, texture and grain — and grain, here, is not a figure of speech. Sedimentary stone has a natural bed, the plane it formed along in the quarry, and it has to go back into the wall on that same orientation. Set the wrong way — face-bedded instead of naturally bedded — a piece of Portland stone or limestone can look perfect on the day it is fitted and begin delaminating within a decade, as the weathering face works against the grain instead of with it. Reading that orientation correctly, in a stone that no longer has an obvious quarry face to reference, is a skill in itself.

The new piece is then dressed by hand to match the tooling of the original — the same chisel and mallet work the surrounding stone carries, not a mechanical saw finish, which reads as obviously modern the moment it weathers differently to everything around it. It is set with a lime-based bedding mortar and pointed with a joint fine enough to read as a shadow line rather than a masonry joint. On an exposed cornice or coping, the top edge is often cut with a dovetail key so the piece cannot work loose under its own weight and the weather it is there to shed.

When a mortar repair is enough, and when it isn't

Not every loss needs an indent. Smaller areas of decay, and sheltered locations where the repair is not carrying the full weight of London weather, are usually better served by a lime-based repair mortar — a plastic repair, built up in layers and finished to the original profile without cutting a socket or introducing a joint at all. It is quicker, less invasive, and on carved or moulded detail it can recover a profile that would be extremely difficult to piece in as a single new stone. We cover the full range of consolidants, mortar repairs and piecing-in on our stone repairs service page.

The line between the two is judgement, not a formula: the extent of the loss, how exposed the location is, whether the stone is load-bearing or purely decorative, and whether there is enough sound substrate left for a mortar repair to key into. It is decided on site, against the specific stone, not from a set of rules applied uniformly across a facade.

Dowelling: the part nobody sees, and the part that decides whether the repair lasts

An indent repair is not held in place by mortar alone. Mortar bonds the joint; it does not carry the mechanical load of a projecting piece of moulding or a step that will be walked on. That job belongs to the dowel — a pin drilled into both the new stone and the sound stone behind it, set with resin, tying the two together mechanically as well as adhesively.

The holes are drilled to align precisely across the joint, cleaned of dust, and the dowel is set with a two-part resin — usually epoxy — that cures fast enough to need the piece held or propped only briefly, and hard enough to transfer load without moving. On a joint that cannot be reached directly, such as a vertical face where the dowel has to be posted into a blind hole, conservators use a technique called mousing: a length of string tied to the dowel guides it into the far hole, then pulls it home once the near piece is offered up. It is a fiddly, unglamorous piece of technique — the kind that never appears in a photograph — and it is used on repairs as significant as major cathedral stonework, because there is no shortcut version that works as well.

What the dowel is made of matters more than almost anything else in the repair. It has to be stainless steel — grade 316, the marine-grade specification, for anything with real exposure — or phosphor bronze. Both are effectively inert against the stone and against water. Neither is what was used for most of the last two centuries.

Why the wrong fixing is worse than no repair at all

Iron cramps and dowels are everywhere in London's Victorian and Edwardian stonework. They were the standard fixing before stainless steel became routine and cheap, and a great deal of that ironwork is still in the walls, quietly corroding. As iron rusts, it expands, and because it is confined inside a drilled hole or a chase cut into solid stone, that expansion has nowhere to go. The stress it generates is enough to crack or shear stone that has stood for a hundred and fifty years without moving — a failure mode conservators call oxide jacking, and not a theoretical one. The same mechanism has cracked masonry at buildings as prominent as St Paul's Cathedral and Winchester Cathedral. It is not a risk confined to minor buildings, and it is not rare.

This is what makes a badly specified repair worse than doing nothing. A mortar patch with no dowel might simply fall out in a decade, which is disappointing but not damaging. A repair pinned with the wrong metal will, on a long enough timeline, split the very stone it was meant to save — and it does this invisibly, from the inside, for years before the crack reaches the surface. By the time it is visible, the damage is usually worse than the original defect the repair was addressing. We have opened up more than one nineteenth or twentieth-century repair on a survey and found exactly this: sound stone, cracked from within, by a fixing nobody thought to question at the time.

What a stone repair survey actually looks for

A repair survey is not just a tally of visible damage. It identifies which losses are structural and which are cosmetic, which stones are sound behind the surface and which are not, and — critically, on a building with any age to it — whether previous repairs used the right materials or are themselves now part of the problem. On a listed building, it also has to establish what a conservation officer will accept: the method statement, the sample repair, and the documentation are as much a part of the job as the stonework itself, and we cover that process in full on our heritage conservation page.

Nothing is specified before that survey happens. The right repair for a given stone depends on how deep the loss goes, what caused it, and what is holding the surrounding material together — none of which is visible from the pavement.

Common questions

What is the difference between a Dutchman repair and an indent repair?

None, in practice — they are the same technique described by two different traditions. In American conservation standards, a new piece of stone let into a damaged area is called a Dutchman repair. In the UK, the same technique is almost always called an indent repair, or piecing-in. Both describe cutting back to sound stone and setting in a matching new piece, dowelled and pointed to sit flush with the surrounding face. We use both terms because clients researching the method, architects especially, often encounter the American term first.

Why does the dowel material matter in a stone repair?

Because the wrong material turns a repair into a future defect. Mild steel and other ferrous fixings corrode, and rust occupies significantly more volume than the metal it replaces. Confined inside a drilled hole in stone, that expansion generates enough force to crack the stone from within — a failure mode called oxide jacking, which has affected buildings as prominent as St Paul's Cathedral. Every dowel we set is stainless steel (grade 316) or phosphor bronze, both effectively immune to this failure.

Can a stone repair be identified once it's finished?

At close range and in raking light, an experienced eye can usually find the joint line — that is by design; a repair should never be so seamless that it misrepresents the building's history. At normal viewing distance, from the pavement or across a courtyard, a well-executed indent repair is very difficult to pick out. The stone is matched in colour, texture and bed orientation, and the surface is hand-dressed to the same tooling as the original rather than left with a mechanical saw finish that would read as obviously modern.

What happens if an old repair used iron fixings instead of stainless steel or bronze?

It becomes the next problem. A huge amount of London's Victorian and Edwardian stonework was originally fixed with iron cramps and dowels, and many twentieth-century repairs used mild steel because it was cheap and available. As that iron corrodes, it expands, and the stone around it cracks or spalls from the inside — often decades after the repair was done, and often mistaken for a new, unrelated defect. Identifying and replacing failed ferrous fixings, not just patching the crack they caused, is a standard part of any repair survey on an older building.